Effective Date: April 1, 2026 | BENED LLC
SIMULATED RESULTS ARE NOT INDICATIVE OF FUTURE PERFORMANCE.
All performance data shown in the Compendium is based on backtested simulations using historical price data. These simulations do NOT account for:
Past performance, whether real or simulated, does not guarantee future results.
This document provides complete transparency about the trading strategies available through TradeCraft's Compendium. We believe users should fully understand the algorithms that will manage their capital before deploying them.
This disclosure covers:
The Compendium is a collection of 30 trading characters, each representing a distinct algorithmic trading strategy. Each character has been simulated against 114+ stock and crypto symbols using 2 years of historical minute-level data (February 2024 – February 2026).
Characters are given memorable persona names (e.g., "Marcus," "Elena," "Viktor") to make strategy selection more intuitive. However, these are algorithmic strategies, not human traders or AI advisors.
Below are the primary strategy types employed by Compendium characters. Each character may use one or more of these approaches.
This strategy identifies oversold conditions using the Relative Strength Index and enters long positions when RSI is low and a reversal candlestick pattern appears.
RS = Average Gain over N periods / Average Loss over N periods
RSI = 100 - (100 / (1 + RS))
Where N = lookback period (typically 14)
1. RSI(14) < 35 (oversold condition)
2. Hammer candlestick pattern detected:
- Lower shadow ≥ 2× body size
- Upper shadow ≤ 0.3× body size
- Candle is bullish (close > open)
3. Price above minimum threshold ($5)
4. Volume above minimum threshold
- Take Profit: 4% gain from entry
- Stop Loss: 2% loss from entry
- Trailing Stop: Activates at 2% profit, trails by 1%
Risks: Mean reversion strategies can suffer significant losses during strong trending markets. An "oversold" stock can continue falling.
Uses exponential moving average crossovers to identify trend changes.
EMA = Price(today) × k + EMA(yesterday) × (1 - k)
Where k = 2 / (N + 1), N = period length
Common periods: 9, 21, 50, 200
BUY: Fast EMA crosses ABOVE Slow EMA
SELL: Fast EMA crosses BELOW Slow EMA
Risks: Crossover strategies generate false signals in choppy, sideways markets. Lag in moving averages can result in late entries and exits.
Uses price deviation from a moving average to identify overbought/oversold conditions.
Middle Band = SMA(20)
Upper Band = SMA(20) + (2 × Standard Deviation)
Lower Band = SMA(20) - (2 × Standard Deviation)
BUY: Price touches or crosses below Lower Band
SELL: Price touches or crosses above Upper Band
Risks: Strong trends can cause prices to "walk the bands," triggering premature entries against the trend.
Moving Average Convergence Divergence identifies momentum shifts.
MACD Line = EMA(12) - EMA(26)
Signal Line = EMA(9) of MACD Line
Histogram = MACD Line - Signal Line
BUY: MACD Line crosses above Signal Line
SELL: MACD Line crosses below Signal Line
Risks: Like all momentum indicators, MACD is a lagging indicator and may miss the best entry points.
Incorporates trading volume as confirmation for price movements.
VWAP = Σ(Price × Volume) / Σ(Volume)
Cumulative throughout the trading session
Risks: Volume patterns that worked historically may not persist. Low-volume stocks may have unreliable VWAP signals.
| Parameter | Value |
|---|---|
| Data Type | 1-minute and 5-minute OHLCV candles |
| Data Range | February 2024 – February 2026 (2 years) |
| Total Candles | 10.1+ million |
| Symbols | 114+ US stocks, ETFs, and cryptocurrencies |
| Source | Licensed market data providers (historical) / Alpaca Markets (live) |
| Parameter | Value |
|---|---|
| Starting Capital | $10,000 / $25,000 / $100,000 |
| Position Sizing | 10% of portfolio per trade |
| Max Concurrent Positions | Varies by strategy |
| Slippage Simulation | Not included |
| Commission Simulation | Not included |
TradeCraft assigns two complementary scores to each strategy+symbol combination. These scores are informational tools to help users understand historical backtest quality — they are not investment advice, recommendations, or predictions of future performance.
A composite metric from 0–100 measuring overall backtest quality across the entire simulation history. Calculated as:
Quality (0-100) = weighted sum of five components:
1. Return / Risk Ratio (30%) — Total return divided by max drawdown
2. Win Rate (20%) — Percentage of profitable trades
3. Drawdown Control (20%) — Penalizes strategies with deep drawdowns
4. Profit Factor (15%) — Gross profit / Gross loss (Bayesian adjusted)
5. Consistency (15%) — Based on trade recency (last_trade_date)
Confidence (0.0 – 1.0) = based on trade count and recency:
- Trade count: scales from 0 → 1.0 as trades approach 100
- Freshness: decays if no trades in recent months
Interpretation:
| Score Range | Label | Meaning |
|---|---|---|
| 75–100 | Elite / PRO | Strong backtest across all metrics with high confidence |
| 60–74 | Good | Solid performance with some areas for improvement |
| 40–59 | Fair | Mixed results — may have strengths offset by weaknesses |
| 0–39 | Weak | Poor backtest results, low confidence, or limited data |
A 30-day trailing score from 0–100 measuring recent performance only. Uses the same Quality × Confidence model as the lifetime score but recalibrated for a shorter time window:
- Window: Last 30 days only (vs. full simulation history)
- Min trades: 3 required (vs. higher threshold for lifetime)
- Full confidence at: 15 trades (vs. 100 for lifetime)
- Return scaling: Higher sensitivity (small 30-day gains score well)
- No freshness decay (all trades are inherently recent)
- Trade frequency replaces consistency (10+ trades/month = full marks)
Interpretation:
| Score Range | Label | Meaning |
|---|---|---|
| 75–100 | Hot | Exceptional recent performance — high activity and returns |
| 60–74 | Active | Strong recent activity with good results |
| 40–59 | Moderate | Some recent activity; mixed or modest results |
| 0–39 | Quiet | Low recent activity or poor recent results |
Comparing the two scores reveals regime sensitivity — whether a strategy+symbol combination is currently performing above or below its historical baseline.
| Indicator | Condition | Interpretation |
|---|---|---|
| 🔥 Heating Up | Momentum exceeds Score by 20+ points | Recent performance significantly outpaces lifetime average. May indicate a favorable market regime for this strategy. Does not predict continuation. |
| 🧊 Cooling Off | Score exceeds Momentum by 20+ points | Recent performance lags behind lifetime average. May indicate an unfavorable current regime despite historically strong results. Does not predict continuation. |
| ⚖️ Steady | Within ±20 points | Recent performance is consistent with historical average. |
These scores are derived entirely from simulated backtest data. They reflect historical patterns, not future outcomes. Specific limitations include:
Scores are informational summaries of backtest history. They are not financial advice.
For complete transparency, the core strategy implementations are available for review upon request. Each character page in the Compendium includes a "Code Verified" badge linking to the exact source file, function, and line numbers implementing that strategy.
This disclosure will be updated whenever:
7.1. When we update this disclosure, you will be notified by email at least 30 days before changes take effect if the changes are material (new risk factors, methodology changes, or removal of strategies).
7.2. A summary of what changed will be provided with each update.
7.3. You will be required to re-acknowledge the updated disclosure before deploying new agents.
If you have questions about any strategy, formula, or implementation detail:
We will respond to strategy-related inquiries within 5 business days.
← Back to Getting Started